The invention relates generally to data communications, and more particularly to a method and apparatus for statistics collection in a data communications network.
Data communications networks include a plurality of nodes, or switches, which are intercoupled to allow users of the network to communicate information to each other. Each of these nodes may maintain a set of statistics relating to the network traffic, where the statistics may be used for purposes such as billing users of the network, evaluating network performance, and isolating faults that may be present within the network. Typically these statistics are maintained via counters internal to the nodes. Such counters are often configured to reset after a predetermined time period, which may be on the order of 15 minutes. As such, the statistics must be collected at least every 15 minutes in order to ensure that no loss of statistic information occurs.
In prior art systems, the statistics collection circuitry for the network typically sent statistics requests using the simple network management protocol (SNMP) or other various proprietary internal protocols appropriate for the nodes within the network. Such statistics requests are of low priority and a significant time period is allocated in which the response to the statistics requests may be received. Unfortunately, if a node is inoperative, or xe2x80x9cdownxe2x80x9d, valuable time and resources are wasted waiting for a response during the lengthy timeout period. In addition, a network is typically only capable of supporting a limited number of statistics polling requests at one time. As such, polling down nodes can waste valuable polling bandwidth, thus jeopardizing the timely collection of all the desirable statistics within the network.
In some prior art systems, a network management platform may detect the down nodes and provide notice to the one or more statistics collectors that may be present within the network. However, the network management platform may be on different workstation than the statistics collectors such that a time delay is experienced for notification of the statistics collectors. Thus, although the notification may prevent some statistics requests to down nodes, the time delay associated with propagation of the notice of the down nodes still presents a certain level of wasted polling bandwidth. In addition, the perspective with which the network management platform perceives the various nodes in the network may differ from that of the statistics collectors that must interface with the nodes in order to collect the statistics for the nodes. As such, nodes may be fully functional from the perspective of the network management platform, whereas a faulty connection may exist between a statistics collector and a node to be polled, thus rendering responses to issued statistics requests to the node impossible.
Therefore, a need exists for a method and apparatus for determining node status prior to statistics request polling that provides timely information regarding the status of nodes from the perspective of the statistics collection circuitry.